fish-shell/src/wcstringutil.rs
Mahmoud Al-Qudsi 8c8da78cf8 Speed up get_case_fold() 5x
Using `c.is_uppercase()` instead of getting the iterator and checking if the
first (and only) lowercase letter of the sequence is the same as the original
input is 5-8x faster (measured via criterion against `/usr/share/dict/words`).

(Additional benefit of forcibly inlining the now iterator-based comparison not
taken into account; this necessitated changing from a closure to a local
function as the inline attribute on closures is not yet supported with the
stable compiler toolchain.)
2024-11-22 16:30:10 -06:00

730 lines
23 KiB
Rust

//! Helper functions for working with wcstring.
use crate::common::{get_ellipsis_char, get_ellipsis_str};
use crate::expand::INTERNAL_SEPARATOR;
use crate::fallback::{fish_wcwidth, wcscasecmp};
use crate::flog::FLOGF;
use crate::libc::MB_CUR_MAX;
use crate::wchar::{decode_byte_from_char, prelude::*};
use crate::wutil::encoding::{wcrtomb, zero_mbstate, AT_LEAST_MB_LEN_MAX};
/// Return the number of newlines in a string.
pub fn count_newlines(s: &wstr) -> usize {
// This is a performance-sensitive function.
// The native filter().count() produces sub-optimal codegen because of overflow checks,
// which we currently enable in release mode. Implement it more efficiently.
let mut count: usize = 0;
for c in s.as_char_slice() {
if *c == '\n' {
count = count.wrapping_add(1);
}
}
count
}
/// Test if a string prefixes another without regard to case. Returns true if a is a prefix of b.
pub fn string_prefixes_string_case_insensitive(proposed_prefix: &wstr, value: &wstr) -> bool {
let prefix_size = proposed_prefix.len();
prefix_size <= value.len() && wcscasecmp(&value[..prefix_size], proposed_prefix).is_eq()
}
/// Test if a string is a suffix of another.
pub fn string_suffixes_string_case_insensitive(proposed_suffix: &wstr, value: &wstr) -> bool {
let suffix_size = proposed_suffix.len();
suffix_size <= value.len()
&& wcscasecmp(&value[value.len() - suffix_size..], proposed_suffix).is_eq()
}
/// Test if a string prefixes another. Returns true if a is a prefix of b.
pub fn string_prefixes_string(proposed_prefix: &wstr, value: &wstr) -> bool {
value.as_slice().starts_with(proposed_prefix.as_slice())
}
/// Test if a string is a suffix of another.
pub fn string_suffixes_string(proposed_suffix: &wstr, value: &wstr) -> bool {
value.as_slice().ends_with(proposed_suffix.as_slice())
}
/// Test if a string matches a subsequence of another.
/// Note subsequence is not substring: "foo" is a subsequence of "follow" for example.
pub fn subsequence_in_string(needle: &wstr, haystack: &wstr) -> bool {
// Impossible if needle is larger than haystack.
if needle.len() > haystack.len() {
return false;
}
if needle.is_empty() {
// Empty strings are considered to be subsequences of everything.
return true;
}
let mut needle_it = needle.chars().peekable();
for c in haystack.chars() {
needle_it.next_if_eq(&c);
if needle_it.peek().is_none() {
return true;
}
}
// We succeeded if we exhausted our sequence.
needle_it.peek().is_none()
}
/// Case-insensitive string search, modeled after std::string::find().
/// \param fuzzy indicates this is being used for fuzzy matching and case insensitivity is
/// expanded to include symbolic characters (#3584).
/// Return the offset of the first case-insensitive matching instance of `needle` within
/// `haystack`, or `string::npos()` if no results were found.
pub fn ifind(haystack: &wstr, needle: &wstr, fuzzy: bool /* = false */) -> Option<usize> {
if needle.is_empty() {
return Some(0);
}
haystack
.as_char_slice()
.windows(needle.len())
.position(|window| {
for (l, r) in window.iter().zip(needle.chars()) {
// In fuzzy matching treat treat `-` and `_` as equal (#3584).
if fuzzy && ['-', '_'].contains(l) && ['-', '_'].contains(&r) {
continue;
}
// TODO Decide what to do for different lengths.
let l = l.to_lowercase();
let r = r.to_lowercase();
for (l, r) in l.zip(r) {
if l != r {
return false;
}
}
}
true
})
}
// The ways one string can contain another.
//
// Note that the order of entries below affects the sort order of completions.
#[derive(Copy, Clone, Debug, Eq, PartialEq)]
pub enum ContainType {
/// Exact match: `foobar` matches `foo`
Exact,
/// Prefix match: `foo` matches `foobar`
Prefix,
/// Substring match: `ooba` matches `foobar`
Substr,
/// Subsequence match: `fbr` matches `foobar`
Subseq,
}
// The case-folding required for the match.
//
// Note that the order of entries below affects the sort order of completions.
#[derive(Copy, Clone, Debug, Eq, PartialEq, PartialOrd, Ord)]
pub enum CaseSensitivity {
/// Exact match: `foobar` only matches `foobar`
Sensitive,
/// Case insensitive match if lowercase input: `foobar` matches `FooBar`.
Smart,
/// Case insensitive: `FooBaR` matches `foobAr`
Insensitive,
}
/// A lightweight value-type describing how closely a string fuzzy-matches another string.
#[derive(Copy, Clone, Debug, Eq, PartialEq)]
pub struct StringFuzzyMatch {
pub typ: ContainType,
pub case_fold: CaseSensitivity,
}
impl StringFuzzyMatch {
pub fn new(typ: ContainType, case_fold: CaseSensitivity) -> Self {
Self { typ, case_fold }
}
// Helper to return an exact match.
#[inline(always)]
pub fn exact_match() -> Self {
Self::new(ContainType::Exact, CaseSensitivity::Sensitive)
}
/// Return whether this is a samecase exact match.
#[inline(always)]
pub fn is_samecase_exact(&self) -> bool {
self.typ == ContainType::Exact && self.case_fold == CaseSensitivity::Sensitive
}
/// Return if we are exact or prefix match.
#[inline(always)]
pub fn is_exact_or_prefix(&self) -> bool {
matches!(self.typ, ContainType::Exact | ContainType::Prefix)
}
// Return if our match requires a full replacement, i.e. is not a strict extension of our
// existing string. This is false only if our case matches and our type is prefix or exact.
#[inline(always)]
pub fn requires_full_replacement(&self) -> bool {
if self.case_fold != CaseSensitivity::Sensitive {
return true;
}
matches!(self.typ, ContainType::Substr | ContainType::Subseq)
}
/// Try creating a fuzzy match for `string` against `match_against`.
/// `string` is something like "foo" and `match_against` is like "FooBar".
/// If `anchor_start` is set, then only exact and prefix matches are permitted.
pub fn try_create(
string: &wstr,
match_against: &wstr,
anchor_start: bool,
) -> Option<StringFuzzyMatch> {
// Helper to lazily compute if case insensitive matches should use icase or smartcase.
// Use icase if the input contains any uppercase characters, smartcase otherwise.
#[inline(always)]
fn get_case_fold(s: &widestring::Utf32Str) -> CaseSensitivity {
if s.chars().any(|c| c.is_uppercase()) {
CaseSensitivity::Insensitive
} else {
CaseSensitivity::Smart
}
}
// A string cannot fuzzy match against a shorter string.
if string.len() > match_against.len() {
return None;
}
// exact samecase
if string == match_against {
return Some(StringFuzzyMatch::new(
ContainType::Exact,
CaseSensitivity::Sensitive,
));
}
// prefix samecase
if match_against.starts_with(string) {
return Some(StringFuzzyMatch::new(
ContainType::Prefix,
CaseSensitivity::Sensitive,
));
}
// exact icase
if wcscasecmp(string, match_against).is_eq() {
return Some(StringFuzzyMatch::new(
ContainType::Exact,
get_case_fold(string),
));
}
// prefix icase
if string_prefixes_string_case_insensitive(string, match_against) {
return Some(StringFuzzyMatch::new(
ContainType::Prefix,
get_case_fold(string),
));
}
// If anchor_start is set, this is as far as we go.
if anchor_start {
return None;
}
// substr samecase
if match_against
.as_char_slice()
.windows(string.len())
.any(|window| wstr::from_char_slice(window) == string)
{
return Some(StringFuzzyMatch::new(
ContainType::Substr,
CaseSensitivity::Sensitive,
));
}
// substr icase
if ifind(match_against, string, true /* fuzzy */).is_some() {
return Some(StringFuzzyMatch::new(
ContainType::Substr,
get_case_fold(string),
));
}
// subseq samecase
if subsequence_in_string(string, match_against) {
return Some(StringFuzzyMatch::new(
ContainType::Subseq,
CaseSensitivity::Sensitive,
));
}
// We do not currently test subseq icase.
None
}
pub fn rank(&self) -> u32 {
// Combine our type and our case fold into a single number, such that better matches are
// smaller. Treat 'exact' types the same as 'prefix' types; this is because we do not
// prefer exact matches to prefix matches when presenting completions to the user.
// Treat smartcase the same as samecase; see #3978.
let effective_type = if self.typ == ContainType::Exact {
ContainType::Prefix
} else {
self.typ
};
let effective_case = if self.case_fold == CaseSensitivity::Smart {
CaseSensitivity::Sensitive
} else {
self.case_fold
};
// Type dominates fold.
effective_type as u32 * 8 + effective_case as u32
}
}
/// Cover over string_fuzzy_match_t::try_create().
pub fn string_fuzzy_match_string(
string: &wstr,
match_against: &wstr,
anchor_start: bool, /* = false */
) -> Option<StringFuzzyMatch> {
StringFuzzyMatch::try_create(string, match_against, anchor_start)
}
/// Implementation of wcs2string that accepts a callback.
/// This invokes `func` with (const char*, size_t) pairs.
/// If `func` returns false, it stops; otherwise it continues.
/// Return false if the callback returned false, otherwise true.
pub fn wcs2string_callback(input: &wstr, mut func: impl FnMut(&[u8]) -> bool) -> bool {
let mut state = zero_mbstate();
let mut converted = [0_u8; AT_LEAST_MB_LEN_MAX];
let is_singlebyte_locale = MB_CUR_MAX() == 1;
for c in input.chars() {
// TODO: this doesn't seem sound.
if c == INTERNAL_SEPARATOR {
// do nothing
} else if let Some(byte) = decode_byte_from_char(c) {
converted[0] = byte;
if !func(&converted[..1]) {
return false;
}
} else if is_singlebyte_locale {
// single-byte locale (C/POSIX/ISO-8859)
// If `c` contains a wide character we emit a question-mark.
converted[0] = u8::try_from(u32::from(c)).unwrap_or(b'?');
if !func(&converted[..1]) {
return false;
}
} else {
converted = [0; AT_LEAST_MB_LEN_MAX];
let len = unsafe {
wcrtomb(
std::ptr::addr_of_mut!(converted[0]).cast(),
c as libc::wchar_t,
&mut state,
)
};
if len == 0_usize.wrapping_sub(1) {
wcs2string_bad_char(c);
state = zero_mbstate();
} else if !func(&converted[..len]) {
return false;
}
}
}
true
}
fn wcs2string_bad_char(c: char) {
FLOGF!(
char_encoding,
L!("Wide character U+%4X has no narrow representation"),
c
);
}
/// Split a string by runs of any of the separator characters provided in `seps`.
/// Note the delimiters are the characters in `seps`, not `seps` itself.
/// `seps` may contain the NUL character.
/// Do not output more than `max_results` results. If we are to output exactly that much,
/// the last output is the remainder of the input, including leading delimiters,
/// except for the first. This is historical behavior.
/// Example: split_string_tok(" a b c ", " ", 3) -> {"a", "b", " c "}
pub fn split_string_tok<'val>(
val: &'val wstr,
seps: &wstr,
max_results: Option<usize>,
) -> Vec<&'val wstr> {
let mut out = vec![];
let val = val.as_char_slice();
let end = val.len();
let mut pos = 0;
let max_results = max_results.unwrap_or(usize::MAX);
while pos < end && out.len() + 1 < max_results {
// Skip leading seps.
match val[pos..].iter().position(|c| !seps.contains(*c)) {
Some(p) => pos += p,
None => {
pos = end;
break;
}
}
// Find next sep.
let next_sep = val[pos..]
.iter()
.position(|c| seps.contains(*c))
.map(|p| pos + p)
.unwrap_or(end);
out.push(wstr::from_char_slice(&val[pos..next_sep]));
// Note we skip exactly one sep here. This is because on the last iteration we retain all
// but the first leading separators. This is historical.
pos = next_sep + 1;
}
if pos < end && max_results > 0 {
assert!(out.len() + 1 == max_results, "Should have split the max");
out.push(wstr::from_char_slice(&val[pos..]));
}
assert!(out.len() <= max_results, "Got too many results");
out
}
/// Joins strings with a separator.
/// This supports both &[&wstr] and &[&WString].
pub fn join_strings<S: AsRef<wstr>>(strs: &[S], sep: char) -> WString {
if strs.is_empty() {
return WString::new();
}
let capacity = strs.iter().fold(0, |acc, s| acc + s.as_ref().len()) + strs.len() - 1;
let mut result = WString::with_capacity(capacity);
for (i, s) in strs.iter().enumerate() {
if i > 0 {
result.push(sep);
}
result.push_utfstr(&s);
}
result
}
pub fn bool_from_string(x: &wstr) -> bool {
matches!(x.chars().next(), Some('Y' | 'T' | 'y' | 't' | '1'))
}
/// Given iterators into a string (forward or reverse), splits the haystack iterators
/// about the needle sequence, up to max times. Inserts splits into the output array.
/// If the iterators are forward, this does the normal thing.
/// If the iterators are backward, this returns reversed strings, in reversed order!
/// If the needle is empty, split on individual elements (characters).
/// Max output entries will be max + 1 (after max splits)
pub fn split_about<'haystack>(
haystack: &'haystack wstr,
needle: &wstr,
max: usize, /*=usize::MAX*/
no_empty: bool, /*=false*/
) -> Vec<&'haystack wstr> {
let mut output = vec![];
let mut remaining = max;
let mut haystack = haystack.as_char_slice();
while remaining > 0 && !haystack.is_empty() {
let split_point = if needle.is_empty() {
// empty needle, we split on individual elements
1
} else {
match haystack
.windows(needle.len())
.position(|window| window == needle.as_char_slice())
{
Some(pos) => pos,
None => break, // not found
}
};
if haystack.len() == split_point {
break;
}
if !no_empty || split_point != 0 {
output.push(wstr::from_char_slice(&haystack[..split_point]));
}
remaining -= 1;
// Need to skip over the needle for the next search note that the needle may be empty.
haystack = &haystack[split_point + needle.len()..];
}
// Trailing component, possibly empty.
if !no_empty || !haystack.is_empty() {
output.push(wstr::from_char_slice(haystack));
}
output
}
#[derive(Eq, PartialEq)]
pub enum EllipsisType {
None,
// Prefer niceness over minimalness
Prettiest,
// Make every character count ($ instead of ...)
Shortest,
}
pub fn truncate(input: &wstr, max_len: usize, etype: Option<EllipsisType>) -> WString {
let etype = etype.unwrap_or(EllipsisType::Prettiest);
if input.len() <= max_len {
return input.to_owned();
}
if etype == EllipsisType::None {
return input[..max_len].to_owned();
}
if etype == EllipsisType::Prettiest {
let ellipsis_str = get_ellipsis_str();
let mut output = input[..max_len - ellipsis_str.len()].to_owned();
output += ellipsis_str;
return output;
}
let mut output = input[..max_len - 1].to_owned();
output.push(get_ellipsis_char());
output
}
pub fn trim(input: WString, any_of: Option<&wstr>) -> WString {
let any_of = any_of.unwrap_or(L!("\t\x0B \r\n"));
let mut result = input;
let Some(suffix) = result.chars().rposition(|c| !any_of.contains(c)) else {
return WString::new();
};
result.truncate(suffix + 1);
let prefix = result
.chars()
.position(|c| !any_of.contains(c))
.expect("Should have one non-trimmed character");
result.split_off(prefix)
}
/// Return the number of escaping backslashes before a character.
/// `idx` may be "one past the end."
pub fn count_preceding_backslashes(text: &wstr, idx: usize) -> usize {
assert!(idx <= text.len(), "Out of bounds");
text.chars()
.take(idx)
.rev()
.take_while(|&c| c == '\\')
.count()
}
/// Support for iterating over a newline-separated string.
pub struct LineIterator<'a> {
// The string we're iterating.
coll: &'a [u8],
// The current location in the iteration.
current: usize,
}
impl<'a> LineIterator<'a> {
pub fn new(coll: &'a [u8]) -> Self {
Self { coll, current: 0 }
}
}
impl<'a> Iterator for LineIterator<'a> {
type Item = &'a [u8];
fn next(&mut self) -> Option<Self::Item> {
if self.current == self.coll.len() {
return None;
}
let newline_or_end = self.coll[self.current..]
.iter()
.position(|b| *b == b'\n')
.map(|pos| self.current + pos)
.unwrap_or(self.coll.len());
let result = &self.coll[self.current..newline_or_end];
self.current = newline_or_end;
// Skip the newline.
if self.current != self.coll.len() {
self.current += 1;
}
Some(result)
}
}
/// Like fish_wcwidth, but returns 0 for characters with no real width instead of -1.
pub fn fish_wcwidth_visible(c: char) -> isize {
if c == '\x08' {
return -1;
}
fish_wcwidth(c).max(0)
}
#[test]
fn test_ifind() {
macro_rules! validate {
($haystack:expr, $needle:expr, $expected:expr) => {
assert_eq!(ifind(L!($haystack), L!($needle), false), $expected);
};
}
validate!("alpha", "alpha", Some(0));
validate!("alphab", "alpha", Some(0));
validate!("alpha", "balpha", None);
validate!("balpha", "alpha", Some(1));
validate!("alphab", "balpha", None);
validate!("balpha", "lPh", Some(2));
validate!("balpha", "Plh", None);
validate!("echo Ö", "ö", Some(5));
}
#[test]
fn test_ifind_fuzzy() {
macro_rules! validate {
($haystack:expr, $needle:expr, $expected:expr) => {
assert_eq!(ifind(L!($haystack), L!($needle), true), $expected);
};
}
validate!("alpha", "alpha", Some(0));
validate!("alphab", "alpha", Some(0));
validate!("alpha-b", "alpha_b", Some(0));
validate!("alpha-_", "alpha_-", Some(0));
validate!("alpha-b", "alpha b", None);
}
#[test]
fn test_fuzzy_match() {
// Check that a string fuzzy match has the expected type and case folding.
macro_rules! validate {
($needle:expr, $haystack:expr, $contain_type:expr, $case_fold:expr) => {
let m = string_fuzzy_match_string(L!($needle), L!($haystack), false).unwrap();
assert_eq!(m.typ, $contain_type);
assert_eq!(m.case_fold, $case_fold);
};
($needle:expr, $haystack:expr, None) => {
assert_eq!(
string_fuzzy_match_string(L!($needle), L!($haystack), false),
None,
);
};
}
validate!("", "", ContainType::Exact, CaseSensitivity::Sensitive);
validate!(
"alpha",
"alpha",
ContainType::Exact,
CaseSensitivity::Sensitive
);
validate!(
"alp",
"alpha",
ContainType::Prefix,
CaseSensitivity::Sensitive
);
validate!("alpha", "AlPhA", ContainType::Exact, CaseSensitivity::Smart);
validate!(
"alpha",
"AlPhA!",
ContainType::Prefix,
CaseSensitivity::Smart
);
validate!(
"ALPHA",
"alpha!",
ContainType::Prefix,
CaseSensitivity::Insensitive
);
validate!(
"ALPHA!",
"alPhA!",
ContainType::Exact,
CaseSensitivity::Insensitive
);
validate!(
"alPh",
"ALPHA!",
ContainType::Prefix,
CaseSensitivity::Insensitive
);
validate!(
"LPH",
"ALPHA!",
ContainType::Substr,
CaseSensitivity::Sensitive
);
validate!("lph", "AlPhA!", ContainType::Substr, CaseSensitivity::Smart);
validate!(
"lPh",
"ALPHA!",
ContainType::Substr,
CaseSensitivity::Insensitive
);
validate!(
"AA",
"ALPHA!",
ContainType::Subseq,
CaseSensitivity::Sensitive
);
// no subseq icase
validate!("lh", "ALPHA!", None);
validate!("BB", "ALPHA!", None);
}
#[test]
fn test_split_string_tok() {
macro_rules! validate {
($val:expr, $seps:expr, $max_len:expr, $expected:expr) => {
assert_eq!(split_string_tok(L!($val), L!($seps), $max_len), $expected,);
};
}
validate!(" hello \t world", " \t\n", None, vec!["hello", "world"]);
validate!(" stuff ", " ", Some(0), vec![] as Vec<&wstr>);
validate!(" stuff ", " ", Some(1), vec![" stuff "]);
validate!(
" hello \t world andstuff ",
" \t\n",
Some(3),
vec!["hello", "world", " andstuff "]
);
// NUL chars are OK.
validate!("hello \x00 world", " \0", None, vec!["hello", "world"]);
}
#[test]
fn test_join_strings() {
let empty: &[&wstr] = &[];
assert_eq!(join_strings(empty, '/'), "");
assert_eq!(join_strings(&[] as &[&wstr], '/'), "");
assert_eq!(join_strings(&[L!("foo")], '/'), "foo");
assert_eq!(
join_strings(&[L!("foo"), L!("bar"), L!("baz")], '/'),
"foo/bar/baz"
);
}
#[test]
fn test_line_iterator() {
let text = b"Alpha\nBeta\nGamma\n\nDelta\n";
let mut lines = vec![];
let iter = LineIterator::new(text);
for line in iter {
lines.push(line);
}
assert_eq!(
lines,
vec![
&b"Alpha"[..],
&b"Beta"[..],
&b"Gamma"[..],
&b""[..],
&b"Delta"[..]
]
);
}
#[test]
fn test_count_newlines() {
assert_eq!(count_newlines(L!("")), 0);
assert_eq!(count_newlines(L!("foo")), 0);
assert_eq!(count_newlines(L!("foo\nbar")), 1);
assert_eq!(count_newlines(L!("foo\nbar\nbaz")), 2);
assert_eq!(count_newlines(L!("\n")), 1);
assert_eq!(count_newlines(L!("\n\n")), 2);
}